IT Management

IT Sense: On the Eve of Disaster

This issue of z/Journal will hit your mailbox well before June 1, a date that has significance in IT shops located in the Gulf Coast and Atlantic Seaboard regions of the U.S. because it signals the start of hurricane season. And from what the guys studying weather at NASA Goddard Space Flight Center and elsewhere are projecting, 2006 promises to be a mean season indeed.

For anyone who didn’t watch TV or the news last year, 2005 was the most active, severe weather season in a very long time. There were a total of 27 named storms—the previous record was 21 in 1933. Of these, 15 became hurricanes (the previous record was 12 in 1969). Of these, four major hurricanes hit the U.S. shores (the previous record was three in 2004), and three were Category 5 Hurricanes on the Saffir-Simpson Hurricane Scale, with winds greater than 155 m.p.h. and storm-related tidal surges generally greater than 18 feet above normal. The most Cat 5 storms making landfall on U.S. shores in recorded history were two each in the 1960 and 1961 seasons.

In parts of the country less frequented by cataclysmic storms, hot weather may be a problem. In the Northeastern corridor (Washington DC, New York, and Boston), there’s already considerable hand-wringing within the electrical power industry around the issue of a potential summer heat wave. The U.S. has the power grid of a third-world country, according to one industry insider, and the capability of our power transmission systems to support the increased demand spawned by hot weather (in the form of more air conditioning, primarily) is dubious at best.

Last year saw brownouts in New York City. The year before saw rolling blackouts. Given NASA’s climate projections this year, which call for hotter days and more of them, the Northeast grid overseers are already commencing a blackout vigil.

So, why the weather report? Simple. Power outages, whether the result of severe weather or transmission shortfalls, are becoming a significant disaster potential for companies—and for IT. The time to prepare is now, not when you find yourself in the dark, hunting for a flashlight.

The alternatives for ensuring your power supply can be separated into two categories: replacement and redundancy. For short-term interruptions in supply, you can probably rely on battery backups or Uninterruptible Power Supplies (UPS). Provided these systems are in good order, they can bridge momentary or short-term (depending on the load and the number of batteries you have) dips and skips in supply. Testing is key, and you should do a test of your backup power system today because load characteristics (how much equipment you need to supply and their collective energy demand) tend to change over time as the business IT infrastructure changes.

For longer term outages, your choices really come down to self-generation (utility replacement)—via diesel generators or other power generation systems—or multiple sourcing (utility redundancy)—connecting to at least two different utility services if you happen to be in an area served by multiple, separate, power service providers. Both alternatives also need to be tested in advance of any disaster, though it may be difficult to project in the case of utility redundancy whether both sources might be disrupted by an interruption that’s geographically wide in scope.

To experienced mainframe data center operators, this stuff may be old hat. To that large cadre of IT professionals who cut their teeth on distributed computing architectures, it may be new information. One thing is for sure: Both groups appeared to be caught on their back heels by the hurricanes that swept through Florida and the Mississippi Delta in 2005.

It’s time to blow the dust off your contingency power plans and test them for their adequacy in keeping mission-critical IT up and running. While you’re at it, you probably need to take inventory of what you thought your key applications and data were when you developed your business continuity/Disaster Recovery (DR) plans way back when (assuming you created such a plan) and to assess whether your original logic still holds true. If gaps have developed over time, you need to plug them now, before another levee breaks.

To aid your efforts and to regain the scent of preparedness planning, you might want to consider attending a free educational event that will take place in my hometown (Tampa, FL) between May 31 and June 1, the official start of the 2006 Hurricane Season. The Disaster Recovery & Data Protection Summit will be held at the Tampa Airport Marriott and will feature knowledgeable speakers, including IT guys who survived last year’s storms, thanks to planning, and free training from the Data Management Institute. Read about it at http://summit.datainstitute.org.

See you at the summit.